JDK-6999606 : runtime/threads/CancellableThreadTest fails with OutOfMemoryError on JDK7 b118 fastdebug bits
  • Type: Bug
  • Component: hotspot
  • Sub-Component: runtime
  • Affected Version: 6
  • Priority: P4
  • Status: Closed
  • Resolution: Not an Issue
  • OS: generic
  • CPU: generic
  • Submitted: 2010-11-12
  • Updated: 2011-04-02
  • Resolved: 2011-04-02
The Version table provides details related to the release that this issue/RFE will be addressed.

Unresolved : Release in which this issue/RFE will be addressed.
Resolved: Release in which this issue/RFE has been resolved.
Fixed : Release in which this issue/RFE has been fixed. The release containing this fix may be available for download as an Early Access Release or a General Availability Release.

To download the current JDK release, click here.
JDK 7
7-poolResolved
Related Reports
Relates :  
JDK-6887571 - Increase default heap config sizes
Description
http://sqeweb.sfbay.sun.com/nfs/results/vm/gtee/HSX/PIT/VM/hs20/b03/jdk7b118/fastdebug/vm/windows-i586/server/comp/windows-i586_vm__server_comp_vm.runtime.testlist/analysis.html

...
Canceller #583 successfully joined CancellableTimer #583 and checks out OK
[2010-11-07T18:15:12.29] Canceller #573 successfully joined CancellableTimer #573 and checks out OK
[2010-11-07T18:15:12.29] Canceller #590 successfully joined CancellableTimer #590 and checks out Exception in thread "Thread-1357" Exception in thread "Thread-1355" Exception in thread "main" java.lang.OutOfMemoryError: unable to create new native thread
[2010-11-07T18:15:12.53] 	at java.lang.Thread.start0(Native Method)
[2010-11-07T18:15:12.53] 	at java.lang.Thread.start(Thread.java:684)
[2010-11-07T18:15:12.53] 	at runtime.threads.CancellableThreadTest.Canceller.run(CancellableThreadTest.java:56)
[2010-11-07T18:15:12.53] 	at java.lang.Thread.run(Thread.java:717)
[2010-11-07T18:15:12.53] java.lang.OutOfMemoryError: unable to create new native thread
[2010-11-07T18:15:12.53] 	at java.lang.Thread.start0(Native Method)
[2010-11-07T18:15:12.53] 	at java.lang.Thread.start(Thread.java:684)
[2010-11-07T18:15:12.53] 	at runtime.threads.CancellableThreadTest.CancellableThreadTest.main(CancellableThreadTest.java:195)
[2010-11-07T18:15:12.53] java.lang.OutOfMemoryError: unable to create new native thread
[2010-11-07T18:15:12.53] 	at java.lang.Thread.start0(Native Method)
[2010-11-07T18:15:12.53] 	at java.lang.Thread.start(Thread.java:684)
[2010-11-07T18:15:12.53] 	at runtime.threads.CancellableThreadTest.Canceller.run(CancellableThreadTest.java:56)
[2010-11-07T18:15:12.53] 	at java.lang.Thread.run(Thread.java:717)
[2010-11-07T18:15:12.53] OK
[2010-11-07T18:15:12.53] Canceller #592 successfully joined CancellableTimer #592 and checks out OK
[2010-11-07T18:15:12.53] Canceller #586 successfully joined CancellableTimer #586 and checks out OK
[2010-11-07T18:15:12.53] Canceller #580 successfully joined CancellableTimer #580 and checks out OK
[2010-11-07T18:15:12.53] Canceller #587 successfully joined CancellableTimer #587 and checks out OK
[2010-11-07T18:15:12.53] VM option '+StartAttachListener'
[2010-11-07T18:15:12.53] # Test level exit status: 1

Note, that test failure occurs on fastdebug bits only. With product bits of b118, the test passes. Note also, that the test passes with b116 both product and fastdebug bits.

An interesting this, is that specifying some small maximum heap size explicitily, e.g., 16m also makes the test pass. Maybe default maximium heap size is set up errorneously too small in the fastdebug version on b118?
The maximum heap size -Xmx only affects the Java heap and this tests has run out of C heap.  Thread stacks are allocated on the C heap.  The thread stack size is 1 or two pages higher in fastdebug/debug mode, but hasn't been changed in a long time.
Of course, I can't reproduce this on my machine but I think I have a lot of memory.  Are there memory quotas on these test machines?  Was this only seen on windows?
Comments
EVALUATION Need more information about machine that can't create 2k threads.
02-03-2011
EVALUATION Ergonomics changes increased the Java heap size, therefore less available system memory for 2k threads to be created on this machine.
17-02-2011